Population structure analysis of C. albicans CF clinical isolates

We recruited 25 antifungal-naive CF patients, 2 adults and 23 pediatric patients (age range 2–22 years), over a period of ~2.5 years (Dec. 2016 to May 2019) from the Béchir Hamza Children’s Hospital in Tunis (BHCHT), a referal center for CF-patient care and treatment in Tunisia. Sputum samples were serially collected as part of the regular clinical examinations performed at the BHCHT and were directly used for the selective isolation of yeast species (S1 Table, see Materials and Methods). In total, 146 C. albicans clinical isolates were recovered from 15 patients out of 25 (S1 Table). Using previously established criteria for chronic colonization by C. albicans [6] (see Materials and Methods), 7 patients were found to chronically carry C. albicans in their airways (patients CF02, CF03, CF04, CF05, CF06, CF07 and CF12, S1 Table).

To better understand the epidemiology and population structure of the CF isolates, we performed multilocus sequence typing (MLST) of 56 strains originating from the 15 CF patients colonized with C. albicans (Fig 1). As vertical transmission of C. albicans from mother to child has been previously shown [15], we also isolated 15 maternal strains from the vaginal and/or oropharyngeal areas of 5 mothers (Mothers of patients CF01, CF02, CF07, CF15 and CF20, S1 Table, Fig 1) to determine if C. albicans strains colonizing the corresponding CF patients could be of maternal origin. Intertestingly, we found that, on many occurrences, individual patients carried strains with identical diploid sequence types (DSTs, see Materials and Methods, Fig 1), suggesting that they were colonized with genetically close strains. Only two patients out of five (CF01 and CF15) carry isolates with DSTs identical to those from their mothers (Fig 1). Notably, ~ half of the patients (8 patients out of 15, CF03, CF05, CF06, CF12, CF15, CF18, CF21 and CF24) carried isolates from clade 4 (Fig 1), including 3 patients (CF03, CF12 and CF15) carrying the same DST (DST95). There is no evidence, so far, pointing to potential strain transmission between these patients; although one could speculate that MoCF15 transmitted the strain to her child (S1 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Dendrogram of the multilocus sequence types of C. albicans strains isolated from the airways of patients with CF.

Phylogenetic relationship of 56 C. albicans clinical isolates recovered from the airways of 15 CF patients (CF01, CF02, CF03, CF04, CF05, CF06, CF07, CF11, CF12, CF15, CF16, CF18, CF20, CF21 and CF23) together with 11 isolates from the mothers of (Mo) patients CF01, CF02, CF07, CF15 and CF20 and reference strain SC5314. Allelic concatenated nucleic acid sequences of the seven loci (MLST, AAT1a, ACC1, ADP1, MPIb, SYA1, VPS13, and ZWF1b) in each strain were phylogenetically analyzed by the Bionumerics 6.0 algorithm (Applied Maths NV, St. Martens-Latem), using the categorical similarity coefficient and the UPGMA clustering method. The allelic profiles were composed of an allele identification number for each gene. An allele combination, known as diploid sequence type (DST), was assigned for the seven loci of each isolate (or obtained for isolates with new DST numbers: MoHBJ, 3581; EJ4-1, 3582; EB1-1, 3583; OM2-1, 3584) according to the C. albicans MLST database. Clade numbers were determined according to the goeBUST algorithm. Seven clades were identified, including clade 4 (green), clade 1 (blue), clade 8 (yellow), clade 3 (orange), clade 17 (pink) and clade 10 (red). Isolate MoFD1-1 clade was not determined, ND (light blue, DST228).

https://doi.org/10.1371/journal.ppat.1012154.g001

To better visualize the extent of similarity/divergence between the identified DSTs, we constructed a minimum-spanning tree (S1 Fig). As expected, DSTs belonging to the same clade clustered together, forming in total 3 clonal complexes and 4 singletons representative of six C. albicans clades (1, 3, 4, 8, 10 and 17) and an unassigned DST (DST228, light blue, S1 Fig). The minimum-spanning tree analysis confirms the prevalence of strains from clade 4 among our set of CF isolates, with a relative predominance of DST95 (S1 Fig). Taken together, our MLST data indicate that each CF patient is colonized with genetically related strains, while a majority of CF patients is colonized with strains from clade 4, a clade enriched with isolates from Middle East Africa [16].

Serial C. albicans clinical isolates from patient CF02 display differential ability to filament and form biofilms

Our findings that C. albicans strains serially recovered from a given CF-patient airway share identifcal DSTs suggest that they are genetically related, which should facilitate comparative analyses between them. Among the 7 CF patients chronically colonized by C. albicans (CF02, CF03, CF04, CF05, CF06, CF07 and CF12, S1 Table), patient CF02 was the most regularly sampled patient within a 12-month period (i.e. 8 times every 1.5 month on average, S1 Table and S2 Fig), yielding a set of 18 isolates sharing DST304 (S2 Fig, strain series HBJ1 to HBJ8 isolated from December 2016 to September 2017). Strains were chronologically numbered, according to their sampling date (S2 Fig). Patient CF02 is a pediatric patient with a presumably dynamically-evolving CF airway environment (S2 Table), as he was co-colonized with C. albicans, P. aeruginosa, S. aureus and Burkholderia cepacia (S1 and S2 Tables). He displayed pulmonary hyperinflation early after birth, a clinical manifestation in CF closely associated with different types of chronic bronchial infections [17]. We reasoned that a systematic phenotypic analysis of the set of 18 serial isolates from patient CF02 can lead to the identification of phenotypic traits associated with the ability of C. albicans to thrive more efficiently within the polymicrobial CF host environment.

We first examined strain fitness of the 18 isolates in liquid YPD and synthetic dextrose (SD) media (Fig 2A). In YPD medium, a slight increase in doubling times of strains HBJ4-3, HBJ6-1 and HBJ6-3 was noticed, compared to the remaining isolates (Fig 2A, top panel). In SD medium, the increased doubling time of strain HBJ6-3 was even more pronounced relative to the remaining strains (Fig 2A, bottom panel). The apparent altered fitness defect of strains HBJ4-3, HBJ6-1 and HBJ6-3 could be due to morphological alterations that interfered with OD_600nm readings, detectable as an irregular growth curve at stationary phase (S3 Fig, red arrows). We therefore examined colony morphology of the complete set of 18 strains on different solid growth media (Fig 2B). On SD medium at 30°C, all strains displayed smooth, equally-sized, creamy patches (Fig 2B). On YP (deprived of glucose), YPD (both at 30°C), RPMI and Spider (both at 37°C), strains HBJ6-1 and HBJ6-3 displayed a strong hairy/fuzzy morphology (Fig 2B). A strong hairy/fuzzy phenotype was also detected in strain HBJ4-3, on RPMI and Spider media at 37°C (Fig 2B). Because filamentous growth and biofilm formation are interconnected processes in C. albicans, we tested the ability of the complete set of strains to form biofilms on fetal bovine serum pre-coated polystyrene microtiter plates, at 37°C (Fig 2C, see Materials and Methods). Strikingly, we found that strains HBJ4-3, HBJ6-1 and HBJ6-3 formed biofilms more efficiently than do the remaining isolates (Fig 2C), correlating with their stronger ability to filament/display altered colony morphology on different growth media (Fig 2B). We conclude that a subset of C. albicans isolates serially recovered from patient CF02 displays enhanced ability to filament and form biofilms.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Phenotypic analyses of the complete set of 18 C. albicans isolates serially recovered from patient CF02.

A. C. albicans isolates were individually grown in rich YPD (top panel) and minimal SD (bottom panel) liquid media in 96-well plates at a starting OD_600nm of 0.1 in 100 μl of YPD or SD at 30°C. OD_600nm was measured every 5 min using a Tecan Infinite 200 reader. Tecan OD_600nm readings were converted into “flask OD600nm” readings using the following formula: OD_Flask = OD_Tecan × 12.2716–1.0543 [95] and doubling times were calculated within the exponential growth interval as previously described [96]. Doubling times (in minutes, average of 3 independent replicates with error bars denoting standard deviations) are indicated on the y-axis for each strain. The average growth rate values of strains HBJ6-3 and HBJ4-3 in YPD (top panel) medium differ from the population by at least one standard deviation from the mean. Doubling time of strain HBJ6-1 in YPD medium (top panel) fell short of reaching this threshold. The red horizontal lines mark the median value. B. Colony morphology phenotypes of the clinical isolates initially patched (5 μL of a cell dilution at an OD_600nm of 0.1) on solid media under filamentation-inducing (RPMI 37°C, Spider 37°C) and -non-inducing (SD at 30°C, YP at 30°C and YPD at 30°C) conditions. C. Biofilm formation was measured three times independently by XTT (2,3-Bis-(2-Methoxy-4-Nitro-5-Sulfophenyl)-2H-Tetrazolium-5-Carboxanilide) assay following growth on FBS-precoated polystyrene microtiter plates in rich medium at 37°C as described in Materials and Methods. The average of three independent replicates are shown as OD_490nm values on the y-axis with error bars denoting standard deviations. Asterisks denote averaged values differing from the population by at least one standard deviation from the mean. The black horizontal line marks the median value. Bottom, photographs of the corresponding two mature biofilms formed independently by the indicated strains on FBS-precoated polystyrene wells, right before quantification by XTT assay.

https://doi.org/10.1371/journal.ppat.1012154.g002

Hyperfilamentous strains display enhanced survival in dual-species biofilms formed with P. aeruginosa

Because strains HBJ4-3, HBJ6-1 and HBJ6-3 displayed a stronger ability to filament and form biofilms, we reasoned that, in the context of CF, these strains could more efficiently survive in mixed biofilms formed with P. aeruginosa. We tested whether the three strains could better compete with P. aeruginosa in dual-species biofilms, than could the poorly filamentous and weak biofilm-former isolate HBJ6-2 (Fig 3). Strain HBJ6-2 was selected in this assay because it was co-isolated with strains HBJ6-1 and HBJ6-3 from the same sputum sample of patient CF02 (S2 Fig). Strain HBJ6-2 is a weakly filamentous isolate, forming pseudohyphae and/or short hyphae in different filamentation-inducing media (S4 Fig, panel A). We performed a well-established C. albicans-P. aeruginosa biofilm assay in microtiter plates at 37°C for 24 h in DMEM medium [14,18,19] (see Materials and Methods). Both HBJ6-2 and the representative hyperfilmentous strain HBJ6-3 could form monospecies biofilms under these conditions (S4 Fig, panel B). We quantified the abundance of HBJ4-3, HBJ6-1, HBJ6-3 and HBJ6-2 cells in dual-species biofilms, relative to their abundance in monospecies biofilms (expressed as % CFUs, Fig 3). To increase clinical relevance, biofilms were grown with P. aeruginosa clinical isolate Pa29575, a strain co-isolated from patient CF02, in addition to P. aeruginosa reference strain PAO1 (Fig 3). We strikingly found that the relative abundance of all hyperfilamentous strains, HBJ6-3, HBJ6-1 and HBJ4-3, in dual-species biofilms formed with either P. aerugiona strains PAO1 or Pa29575 was significantly increased compared to that of poorly filamentous isolate HBJ6-2 (Fig 3), indicating that although the four strains are genetically similar, they differ in their capacity to survive in dual-species biofilms formed with P. aeruginosa.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Quantification of the relative survival of C. albicans CF isolates HBJ6-2, HBJ6-3, HBJ6-1 and HBJ4-3 in dual species biofilms formed with P. aeruginosa.

Preformed 24-h biofilms with either C. albicans alone (Ca, monospecies biofilms made by HBJ6-2, HBJ6-3, HBJ6-1 and HBJ4-3) or both C. albicans and P. aeruginosa (dual-species biofilms, Ca + PAO1, Ca + Pa29575) were treated with DNase I to degrade the extracellular matrix then detached from the surface of wells by scraping (see Materials and Methods). The resulting cell suspensions were diluted and plated onto antibiotic-containing YPD agar to determine the percent CFUs (y-axis, %CFUs) of C. albicans cells co-cultured with P. aeruginosa PAO1 or Pa29575 relative to biofilm growth of C. albicans cells alone (y-axis, %CFUs set to 100%). The experiments were carried out 3 to 5 times independently and data are plotted as the average of 5 independent biological replicates on the y-axis with error bars denoting standard deviations. Statistical analysis was performed using the non-parametric Mann-Whitney test that compares %CFUs for HBJ6-3, HBJ6-1 or HBJ4-3 in each co-incubation condition (Ca + PAO1 or Ca + Pa29575) with the %CFUs for HBJ6-2 in the corresponding co-incubation condition. *, P<0.05; **, P<0.01.

https://doi.org/10.1371/journal.ppat.1012154.g003

Whole-genome sequencing of serial C. albicans clinical isolates from patient CF02 confirms their genetic relatedness

Although MLST is a powerful tool for analyzing the population structure of C. albicans isolates from multiple host sources and niches of the human body [16] (Fig 1), it does not allow to perform high-resolution and/or in-depth analyses of the genomic alterations that could have occurred within a set of serial isolates from a given patient, which could point to adaptive changes within the host environment. Hence, we subjected the complete set of HBJ isolates from patient CF02 to whole-genome sequencing (WGS) and mapped the resulting reads to the diploid Assembly 22 of the C. albicans reference genome (strain SC5314, see Materials and Methods). The average sequencing depth was 138×, with a range of 101× to 169×, allowing to robustly determine the extent of genetic variability in our set of isolates. We performed variant calling and filtering analyses, to identify high-confidence single nucleotide polymorphisms (SNPs) and determine the genetic relatedness of the isolates (see Materials and Methods). Compared to the SC5314 reference genome, we found 132,586 highly confident SNPs across the 18 HBJ isolates (S4 Table). Analysis of genetic variation revealed that each of the 18 isolates carried on average 102,148 heterozygous SNPs (versus 65,629 heterozygous SNPs on average in each of 182 clinical isolates collected worldwide [20]), indicating that their genomes were extensively heterozygous (S5 Fig). Mapping of the heterozygous SNPs using a 10-kb sliding window confirmed the genetic relatedness of the 18 serial isolates (S5 Fig), differing from each other by 1,337 SNPs on average (S5 Table). We constructed a phylogenetic tree where they were compared to a set of 182 isolates collected worldwide by Ropars et al. [20] (see Materials and Methods, Fig 4). We found that the 18 strains strongly clustered with each other among members of clade 10 (Fig 4, red cluster), displaying a highly clonal subpopulation structure (Fig 4, indicated with asterisks). We verified if copy number variations may have occurred in their genomes, by determining the allele balance at heterozygous sites (ABHet, see Materials and Methods) and plotting ABHet values across the 8 chromosomes (S6 Fig). No copy-number variations were detected in any of the 18 strains (S6 Fig). Taken together, our WGS data confirm the genetic relatedness of the 18 serial isolates recovered from patient CF02 and validate our assumption of their clonal origin.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Maximum likelihood phylogenetic tree showing relationships between 200 C. albicans isolates including those from patient CF02.

A total of 200 C. albicans isolates (182 isolates from Ropars et al. together with 18 isolates used in this study, shown in red and indicated by asterisks) were clustered into 17 distinct genetic clades, including 12 clades previously found using multilocus sequencing typing (MLST) and five recent ones (A, B, C, D, and E) described by Ropras et al. [20]. The majority of the isolates belonged to clades 1 (n  =  40), 4 (n  =  27), 10 (n = 21), and 13 (n = 35). C. albicans isolates from patient CF02 (n = 18) were clustered in clade 10 (indicated by asterisks), with a clonal subpopulation structure. Scale bar, 0.1 nucletide substitutions per site on average.

https://doi.org/10.1371/journal.ppat.1012154.g004

A handful number of non-synonymous SNPs differentiate strains HBJ4-3, HBJ6-1 and HBJ6-3 from the remaining CF02 isolates

Although genetically related to each other, the set of 18 serial isolates includes 3 strains (HBJ4-3, HBJ6-1 and HBJ6-3) that phenotypically differ from the remaining isolates. First, the three strains are hyperfilamentous (Fig 2B) and are potent biofilm formers (Fig 2C). Second, they better compete with P. aeruginosa in dual-species biofilms, compared to the poorly filamentous and weak biofilm-former strain HBJ6-2 (Fig 3). We hypothesize that these strains acquired genetic alterations that led to the emergence of new traits allowing them to better compete with other CF-associated microbes, such as P. aeruginosa, and better survive in the CF-lung environment. Consequently, we mined the list of 132,586 highly confident SNPs across the 18 HBJ isolates (S4 Table) to identify SNPs that are exclusively shared by the three strong biofilm-former strains HBJ4-3, HBJ6-1 and HBJ6-3 (S6 Table). We found 239 SNPs respecting this criterion, including 154 in protein-coding regions (ORFs and exons), 76 in intergenic regions, 6 within introns and 3 in the major repeat sequence MRS-1 (S6 Table). Among the 154 SNPs in protein-coding sequences, we identified 34 non-synonymous SNPs in 19 ORFs (S7 Table). We reasoned that if non-synonymous SNPs were potentially linked to adaptive growth in the CF lung environment, they should be rarely (if not at all) detected among strains isolated from niches other than the CF airways. We therefore excluded 20 non-synonymous SNPs prevalent among 182 non-CF isolates previously sequenced by Ropars et al. [20], narrowing down the list to 14 non-synonymous SNPs in 7 ORFs (Table 1). Interestingly, all 14 non-synonymous SNPs were heterozygous SNPs; half of them were located in orf19.4346, encoding a homolog of S. cerevisiae Sec16, involved in vesicular trafficking, while two were carried by orf19.427, encoding a homolog of S. cerevisiae telomeric DNA binding protein Rif1 (Table 1). The remaining 5 non-synonymous SNPs were individually located in 3 uncharacterized ORFs (orf19.415, orf19.2310 and orf19.4406) and 2 ORFs encoding transcription factors of the zinc cluster family, Rob1 and Zfu3 (Table 1). Rob1 is a positive regulator of biofilm formation and filamentous growth in C. albicans [21–23], whereas Zfu3 is an uncharacterized transcription factor that was previously shown to negatively regulate C. albicans morphogenesis [24,25]. In conclusion, our analyses identified a handful number of non-synonymous heterozygous SNPs in hyperfilamentous and strong biofilm-former strains HBJ4-3, HBJ6-1 and HBJ6-3 that differentiates them from the remaining CF02 isolates and from 182 non-CF strains collected worldwide. We postulate that a subset of these SNPs could have mediated the emergence of CF-adaptive traits in these isolates.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. List of 14 non-synonymous SNPs found in 7 ORFs that are exclusively shared between hyperfilamentous and strong biofilm former strains HBJ4-3, HBJ6-1 and HBJ6-3, and are not present across 182 non-CF strains previously sequenced by Ropars et al. [20].

https://doi.org/10.1371/journal.ppat.1012154.t001

A single amino-acid change, G299E, in Rob1 is sufficient to drive hyperfilamentation and increased biofilm formation

Our examination of the list of non-synonymous SNPs exclusively shared by the three hyperfilamentous and strong biofilm-former strains HBJ4-3, HBJ6-1 and HBJ6-3 (Table 1) drew our attention to those carried by the transcription factor-encoding genes ROB1 and ZFU3 (Table 1). We focused on these two genes for three reasons. First, they encode decision-making proteins (i.e. transcription factors), that enable an extensive reprogramming of the transcriptome to induce major phenotypic and developmental changes [26]. Second, zinc cluster transcription fators are drivers of evolutionary adaptation in fungi, by acquiring gain-of-function mutations that circumvent environmental pressure [27]. Third, while Rob1 was shown to be part of a tightly-knit transcriptional network that controls biofilm development [22], both Rob1 and Zfu3 were shown to regulate C. albicans filamentous growth in a haploinsufficient manner [24]. We hypothesized that the two identified SNPs in ROB1 (G to A substitution at position 962 relative to ATG translation start site of the ROB1 ORF) and ZFU3 (G to A substitution at position 489 relative to ATG translation start site of the ZFU3 ORF), which respectively introduce amino acid changes at positions 299 (Glycine to glutamic acid, G299E) and 127 (Alanine to threonine, A127T) in their deduced protein sequences (Table 1), could be responsible–at least in part–for the acquisition of the hyperfilamentation and strong biofilm formation phenotypes in clinical isolates HBJ4-3, HBJ6-1 and HBJ6-3.

We used site-directed mutagenesis to introduce the G>A substitution at positions 962 and 489 relative to the ATG translation start site of the ROB1 and ZFU3 ORFs, respectively, in the poorly filamentous and weak biofilm-former strain HBJ6-2 (see Materials and Methods). We generated two pairs of allele replacement cassettes, one pair with the dominant selection marker SAT1 (conferring resistance to nourseothricin) carrying a mutated (ROB1^G299E) or an equivalent wild-type version of ROB1 (serving as a control); and another pair with the dominant selection marker hygB (conferring resistance to hygromycin) harboring a mutated (ZFU3^A127T) or an equivalent wild-type version of ZFU3 (see Materials and Methods). The four cassettes were used to (co)transform clinical isolate HBJ6-2 and obtain independent single and double heterozygous mutants (together with matching wild-type controls, see Materials and Methods). We tested the ability of the resulting mutants to alter colony morphology, cellular filamentation and biofilm formation (Fig 5). On solid RPMI medium at 37°C, the introduction of the G299E mutation in Rob1 was sufficient to strongly induce a fuzzy phenotype in the resulting heterozygous mutant, phenocopying the hyperfilamentous HBJ6-3 isolate (Fig 5A). Introduction of the A217T mutation in Zfu3, did not affect colony morphology, phenocopying the parental HBJ6-2 strain as well as the control strains carrying wild-type alleles of ROB1 or ZFU3 (Fig 5A). We examined the morphology of the heterozygous double mutant and found that it phenocopied both single ROB1^G299E mutant and strain HBJ6-3 (Fig 5A). Equivalent results were obtained on the filamentation-inducing solid Spider medium at 37°C, albeit with a slightly stronger wrinkling phenotype displayed by strain HBJ6-3 (Fig 5A). We also verified cellular filamentation in liquid Spider medium at 37°C (Fig 5B). While both heterozygous single ROB1^G299E and double mutants displayed a marked filamentous growth, the ZFU3^A217T heterozygous mutant failed to do so (Fig 5B), further supporting no role for the ZFU3^A217T mutation in filamentous growth, at least in a heterozygous context.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Impact of ROB1^G299E and ZFU3^A217T mutations on C. albicans morphogenesis and biofilm formation.

A. Representative colony morphology (left and middle columns) and biofilm formation (right column) phenotypes of HBJ6-2 heterozygous mutant derivatives (3 independent clones were generated for each mutant, see S8 Table) carrying the ROB1-SAT1 (+ ROB1-SAT1, control strain, clone # RCC13, S8 Table), the ROB1^G299E-SAT1 (+ROB1^G299E-SAT1, clone # RC10, S8 Table), the ZFU3-hygB (+ ZFU3-hygB, control strain, clone # ZCC5, S8 Table), the ZFU3^A217T-hygB (+ ZFU3^A217T-hygB, clone # ZC6, S8 Table) and both ROB1^G299E-SAT1 and ZFU3^A217T-hygB (+ ROB1^G299E-SAT1 + ZFU3^A217T-hygB, double mutant, clone # 2MC7, S8 Table) allele replacement cassettes are shown together with those of the parental poorly filamentous HBJ6-2 (Parental HBJ6-2, top row) and the hyperfilamentous HBJ6-3 (bottom row) strains. Strains were patched (5 μl of a cell dilution at an OD_600nm of 0.1) on solid media under filamentation-inducing (RPMI, left column; Spider, middle column) conditions and grown at 37°C for 5 days. The indicated strains were also induced to form biofilms for 24 h at 37°C on FBS-precoated polystyrene wells (right columns) as described in panel C. B. The strains tested on solid Spider medium in panel A were also tested in liquid Spider medium at 37°C (indicated on top of each image). Overnight, saturated, pre-cultures in YPD medium were diluted to an OD_600nm of 0.3 in 2 mL of liquid Spider medium in 12-well polystyrene plates. The diluted cultures were incubated at 37°C under vigorous shaking for 4 hours. The morphology of C. albicans cells from each culture was examined with a light microscope, at 40× magnification. Scale bar, 100 μm. C. Quantitative biofilm formation assay with the parental strain HBJ6-2 (Parental HBJ6-2), the control strain carrying the ROB1-SAT1 allele replacement cassette (+ ROB1-SAT1), the single ROB1^G299E mutant (+ROB1^G299E-SAT1, clone # RC10, S8 Table), the ROB1^G299E ZFU3^A217T double mutant (+ ROB1^G299E-SAT1 + ZFU3^A217T-hygB, clone # 2MC7, S8 Table) and the hyperfilamentous and strong biofilm-former strain HBJ6-3 (HBJ6-3) was performed 5 times independently on FBS-precoated polystyrene microtiter plates in YPD medium at 37°C. An initial incubation for 30 min allowed adherence of cells to the FBS-precoated polystyrene surface, followed by washing and static re-growth in YPD medium at 37°C for 24 h. After a final washing step, biofilm formation was assessed by quantification of biofilm density using spectrophotometry (see Materials and Methods). Averaged OD_600nm values from the 5 independent biological replicates are shown on the y-axis, with error bars denoting standard deviations. Statistical analysis was performed using the non-parametric Mann-Whitney test. ns, non-significant; **, P<0.01.

https://doi.org/10.1371/journal.ppat.1012154.g005

Based on the conclusive evidence that the ROB1^G299E mutation enhances filamentous growth in the poorly filamentous HBJ6-2 strain, we tested its impact on biofilm development (Fig 5, panels A and C). We strikingly found that the heterozygous single ROB1^G299E mutant developed into a dense biofilm layer on the surface of microtiter-plate wells, phenocopying the strong biofilm-former strain HBJ6-3 (Fig 5A). Again, the heterozygous double mutant was indistinguishable from the single ROB1^G299E mutant (Fig 5A), reinforcing no contribution of the ZFU3^A217T heterozygous mutation in biofilm formation. Our results were confirmed with a quantitative biofilm assay, assessing the optical density of biofilm layer at 600 nm (see Materials and Methods, Fig 5C). Both heterozygous single ROB1^G299E and double mutants strongly formed biofilms at similar levels (Fig 5C). Knowing the function of Rob1 as an activator of filamentous growth and biofilm formation [21–24], we conclude that the G299E substitution in Rob1 acts as a phenotypic gain-of-function mutation, sufficient, in a heterozygous context, to enhance filamentous growth and biofilm development in C. albicans.

The G299E mutation in Rob1 induces the expression of Rob1 targets

The phenotypic consequences of the G299E substitution in Rob1 suggest that this mutation may convert transcription factor Rob1 into a hyperactive state, hence impacting on transcriptional regulation of Rob1 targets by increasing their expression in the absence of inducing cues, as previously shown with other regulators of this family [26,28]. We tested this hypothesis in HBJ6-2 strain derivatives ROB1^G299E-SAT1/ROB1 and ROB1-SAT1/ROB1—which genetically differ only by the G299E gain-of-function mutation in Rob1 (Fig 6). Both strains were grown in YPD medium at 30°C, then total RNA was extracted and subjected to q/RT-PCR analyses (see Materials and Methods). We qualitatively (Fig 6A) and quantitatively (Fig 6B) assessed the expression levels of Rob1 targets ALS3, HWP1 and ECE1 [22,29], as well as ROB1 itself, using specific primers (S3 Table). While ALS3, HWP1 and ECE1 gene expression levels in the HBJ6-2 derivative ROB1-SAT1/ROB1 were undetectable by agarose gel electrophoresis of RT-PCR products, the three genes displayed increased expression levels in the ROB1^G299E-SAT1/ROB1 mutant (Fig 6A). ROB1 gene expression was detectable at comparable levels in both HBJ6-2 strain derivatives, paralleling with the expression patterns of ACT1 and TEF3 loading controls (Fig 6A). Quantitative RT-PCR analyses further confirmed the induction of ALS3, HWP1 and ECE1 gene-expression in strain ROB1^G299E-SAT1/ROB1 compared to strain ROB1-SAT1/ROB1 (Fig 6B), reaching levels as high as 4.7-fold, 8.1-fold and 35.8-fold, respectively (Fig 6B). Taken together, these data demonstrate that the G299E substitution in Rob1 acts as a gain-of-function mutation, leading to strong activation of the expression of Rob1 targets.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Impact of the ROB1^G299E mutation on the expression of Rob1 targets.

A. The expression of ROB1, ECE1, HWP1 and ALS3 in HBJ6-2 strain derivatives ROB1/ROB1-SAT1 (left) and ROB1/ROB1^G299E-SAT1 (right) was qualitatively assessed by reverse-transcription polymerase chain reaction (RT-PCR) and agarose gel electrophoresis of the RT-PCR products using specific primers listed in S3 Table. The ACT1 and TEF3 loading control amplification signals are shown at the bottom of the panel. B. The relative expression of the same genes (y-axis, relative expression levels (n-fold)) was also quantitatively analyzed by qRT-PCR. The expression of endogenous gene ACT1 was used as the normalization standard, and the relative expression of the indicated genes (bottom) in the ROB1^G299E-SAT1/ROB1 mutant (light gray-filled histograms) compared to its expression in the ROB1-SAT1/ROB1 (open histograms) was determined using the cycle threshold (ΔΔCt) method using the average ΔCt values of the ACT1 gene in the ROB1-SAT1/ROB1 strain as a calibrator (see Materials and Methods). The assays were carried out three times independently and statistical analysis was performed using a Welch’s t-test. ns, not significant; *, P<0.05; **, P<0.01.

https://doi.org/10.1371/journal.ppat.1012154.g006

Loss-of-function of ROB1 abolishes the hyperfilamentation and strong biofilm formation phenotypes in strain HBJ6-3

Our finding that the gain-of-function mutation G299E in Rob1 converts the poorly filamentous and weak biofilm-former strain HBJ6-2 into a hyperfilamentous and strong biofilm-former mutant fostered us to test whether deletion of ROB1 in strain HBJ6-3 would lead to the reverse phenotype. It would also assess the contribution of ROB1 in the ability of HBJ6-3 to hyperfilament and strongly develop into biofilms. We used the transient CRISPR-Cas9-mediated gene deletion strategy [30] to disrupt both copies of ROB1 in clinical isolates HBJ6-2 and HBJ6-3 (see Materials and Methods, Fig 7). In the poorly filamentous isolate HBJ6-2, disruption of both copies of ROB1 converted the weak colony wrinkling phenotype observed on RPMI medium, at 37°C, into a smooth colony phenotype in the resulting mutant (Fig 7A, left column). Disruption of ROB1 in strain HBJ6-3, however, had a dramatic impact on colony morphology; with total loss of the strong colony fuzziness/wrinkling on filamentation-inducing solid media RPMI and Spider, at 37°C (Fig 7A, left and middle columns). The dense biofilm layer formed by strain HBJ6-3 was totally lost in the resulting rob1Δ/rob1Δ mutant (Fig 7A, right column). A quantitative biofilm assay confirmed our observation (Fig 7B), showing that biofilm density levels of strain HBJ6-3 significantly decreased in the resulting mutant, reaching levels as low as those detectable in both the HBJ6-2 strain and the rob1Δ/rob1Δ derivative (Fig 7B). Taken together, these data, along with data from Fig 5 indicate that the ROB1 locus is the genetic driver of the ability of CF02 clinical isolates HBJ4-3, HBJ6-1 and HBJ6-3 to hyperfilament and strongly form biofilms, due to the acquisition of a heterozygous gain-of-function mutation in ROB1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Contribution of ROB1 to colony morphology and biofilm formation in strain HBJ6-3.

A. Representative colony morphology (left and middle columns) and biofilm formation (right column) phenotypes of HBJ6-2- and HBJ6-3-derived homozygous rob1Δ/rob1Δ mutants (5 independent clones were generated from each parental strain, see S8 Table) that were subjected to a CRISPR-Cas9-driven homozygous inactivation of the ROB1 gene using the dominant selection marker SAT1 (see Materials and Methods) are shown. The parental poorly filamentous HBJ6-2 strain (ROB1/ROB1, top row) and the rob1Δ/rob1Δ mutant derivative (2^nd row from top, rob1Δ/rob1Δ, strain MGY16, clone # KOC28, S8 Table), together with the parental hyperfilamentous strain HBJ6-3 (ROB1^G299E/ROB1, third row from top) and the rob1Δ/rob1Δ mutant derivative (bottom row, rob1Δ/rob1Δ, strain MGY17, clone # KOC7, S8 Table) were patched (5 μl of a cell dilution at an OD_600nm of 0.1) on solid media under filamentation-inducing (RPMI, left column; Spider, middle column) conditions and grown at 37°C for 5 days. The indicated parental strains and their mutant derivatives were also induced to form biofilms for 24 h at 37°C on FBS-precoated polystyrene wells (right columns) as described in panel B. B. Quantitative biofilm formation assay was performed 6 times independently on FBS-precoated polystyrene microtiter plates in YPD medium at 37°C with the parental strains and the corresponding rob1Δ/rob1Δ mutant derivatives described in panel A (shown beneath the x-axis). An initial incubation for 30 min allowed adherence of cells to the FBS-precoated polystyrene surface, followed by washing and static re-growth in YPD medium at 37°C for 24 h. After a final washing step, biofilm formation was assessed by quantification of biofilm density using spectrophotometry (see Materials and Methods). Averaged OD_600nm values from the 6 independent biological replicates are shown the y-axis, with error bars denoting standard deviations. Statistical analysis was performed using the non-parametric Mann-Whitney test. ns, non-significant; ***, P<0.001.

https://doi.org/10.1371/journal.ppat.1012154.g007

The ROB1^G299E mutation is sufficient to confer increased survival of C. albicans in dual-species biofilms formed with P. aeruginosa

Because the hyperfilamentous and strong biofilm-former strains HBJ4-3, HBJ6-1 HBJ6-3 were able to better compete with P. aeruginosa in dual-species biofilms than was the poorly filamentous and weak biofilm-former strain HBJ6-2 (Fig 3), we hypothesized that the ROB1^G299E mutation conferred such an advantage to these isolates. We therefore challenged both HBJ6-2 strain derivatives ROB1^G299E-SAT1/ROB1 and ROB1-SAT1/ROB1—which genetically differ only by the G299E gain-of-function mutation in Rob1—with P. aeruginosa strains PAO1 (reference lab strain) and Pa29575 (clinical isolate from patient CF02) in a dual-species biofilm assay (Fig 8). We compared the relative survival of both mutants as described above (Fig 3, see Materials and Methods). We found that the relative abundance of the ROB1^G299E-SAT1/ROB1 mutant in dual-species biofilms formed with either P. aerugiona strains PAO1 or Pa29575 was significantly increased compared to that of the ROB1-SAT1/ROB1 mutant control strain (Fig 8). Our results demonstrate that the ROB1^G299E mutation is sufficient to confer increased survival of C. albicans in dual-species biofilms formed with P. aeruginosa, reflecting its clinical relevance in the context of CF.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Impact of ROB1^G299E mutation on survival of C. albicans in dual-species biofilms formed with P. aeruginosa.

Preformed 24-h biofilms with either C. albicans alone (monospecies biofilms, Ca) or both C. albicans and P. aeruginosa (dual-species biofilms, Ca + PAO1, Ca + Pa29575) were treated with DNase I to degrade the extracellular matrix then detached from the surface of wells by scraping (see Materials and Methods). The resulting cell suspensions were diluted and plated onto antibiotic-containing YPD agar to determine the percent CFUs (y-axis, %CFUs) of C. albicans cells (the HBJ6-2-derived heterozygous control strain ROB1/ROB1-SAT1 # RCC13 or the ROB1/ROB1^G299E-SAT1 mutant # RC10, S8 Table) co-cultured with P. aeruginosa PAO1 or Pa29575 relative to biofilm growth of C. albicans cells alone (y-axis, %CFUs set to 100%). The experiments were carried out three times independently and data are plotted as the average of three independent biological replicates on the y-axis with error bars denoting standard deviations. Statistical analysis was performed using a Student’s t test. **, P<0.01; ***, P<0.001.

https://doi.org/10.1371/journal.ppat.1012154.g008

Ethics statement

All procedures complied with ethical standards for human investigations and the principles of the Declaration of Helsinki. Written informed consent for minors was obtained from the patient parent or legal guardian. In case of sampling of family relatives or CF patients aged ≥18 years, an informed consent was signed. The present study protocol was reviewed and approved by the Ethics Committee at Institut Pasteur de Tunis (Application file # 2019/22/I/ LR16IPT/V3).

CF-patient clinical data and sputum sample collection

Twenty-five antifungal-naive CF patients were recruited in this study (S1 Table). Confirmed CF diagnosis was established based on clinical symptoms with two positive sweat chloride tests, molecular analysis of the CFTR gene and identification of the CF-causing mutation (S2 Table). Data collected includes patient age, gender, CFTR mutation, body mass index, chronic lung colonization with P. aeruginosa, mucoid phenotype of P. aeruginosa, allergic bronchopulmonary aspergillosis, bronchiectasis, the presence of B. cepacia, exacerbation frequency and hepatic cirrhosis (S1 and S2 Tables). Sputum samples were collected as part of routine clinical care from all CF patients, with the assistance of a medical practitioner. In total, 119 sputum samples and one tongue swab sample were obtained over a period of ~2.5 years (December 2016 to April 2019, S1 Table). The studied population was comprised of 23 pediatric and 2 adult patients, with an age range encompassing 2–22 years, the median age being 5 years (S1 Table). The number of samples per patient depended on the frequency of the visits to the BHCHT, Tunisia (S1 Table). We also collected samples from 19 mothers of CF patients, including 19 oropharyngeal swab samples and 3 sputum samples to test for the presence of C. albicans and assess potential mother-to-child transmission (S1 Table).

Microbiological method for the isolation and identification of Candida spp

The collection of sputum specimens was performed at Children’s Hospital in Tunis. To prevent microbiological contamination of the expectorated sputa, all patients were asked to rinse their mouth twice with water then cough in order to expectorate sputum into a sterile container. Collected specimens from subjects were solubilized by dilution into normal saline (0.9% NaCl), then plated onto sabouraud dextrose agar, containing chloramphenicol (100 μg/ml) to inhibit the growth of bacteria and cycloheximide (50 μg/ml) to inhibit the growth of saprophytic and growth medium-contaminating fungi, then incubated at 30°C for one week. The isolated yeast species were tested for their ability to form germ tubes and chlamydospores. Only isolates capable of chlamydospore formation were streaked onto chromagar (chromID® Candida biomérieux, France) and incubated at 37°C for 5 days. Single blueish colonies (C. albicans or C. dublinensis) were restreaked on YPD agar medium then subjected to species identification using two independent approaches: i) Internal transcribed spacer 2 (ITS2) amplification using primer pair 3271-ITS2F and 3271-ITS2R [92] (S3 Table) followed by sequencing with an ABI prism 3500 genetic analyzer (Applied Biosystems) and/or 2) Matrix-Assisted Laser Desorption/Ionization-Time of Flight (MALDI-ToF) biotyper at hôpital Necker-Enfants malades in Paris, France (S1 Table). In total, 146 C. albicans strains were isolated from 15 patients (S1 Table). Patients with three or more consecutive growth of C. albicans in sputum cultures or in ≥ 50% of cultures obtained within a 12-month period were considered to have chronic C. albicans colonization, those with positive cultures in 25–49% of sputum samples obtained within 12 months were classified as having intermittent colonization, and those with no positive cultures or positive findings in < 25% of cultures obtained over 12 months were defined as not colonized [6]. Glycerol stocks were prepared for each identified isolate and stored at -80 C.

Genotyping of C. albicans clinical isolates by Multilocus Sequence Typing

MLST of C. albicans isolates involves the PCR amplification and DNA sequence analysis of 300–400 bp regions from seven housekeeping genes (AAT1a, ACC1, ADP1, PMI1b, SYA1, VPS13, and ZWF1b) that are under stabilising selection pressure [93]. MLST data are represented by genotype numbers, which define unique sequences for pairs of alleles, and diploid sequence types (DSTs), which define unique combinations of genotypes. Total genomic DNA of C. albicans clinical isolates was extracted using a rapid miniprep protocol described previously by Liu et al. [94]. The extracted DNA was subjected to PCR amplification of the seven housekeeping genes with primers listed in S3 Table. The nucleotide polymorphisms of the resulting sequences were used for MLST analysis as described by Bougnoux et al. [93]. Shortly, PCR was carried out in 50-μl reaction volume containing 2 μl of genomic DNA (10 ng/μl), 1 μl of each primer (0.2 μM), 1 unit of Taq DNA polymerase (Thermo Fisher Scientific), 5 μl of 10X reaction buffer, 5 μl MgCl₂ (2.5mM) and 4 μL of deoxynucleoside triphosphate (2mM). The PCR products of the seven housekeeping genes were sequenced using the ABI PRISM BigDye Terminator v3.1 sequencing kit according to the manufacturer’s instructions, then runned on an ABI prism 3500 genetic analyzer sequencer (Applied Biosystems). The resulting FASTA files were uploaded to the C. albicans MLST database (https://pubmlst.org/organisms/candida-albicans) for DST identification.

Phylogenetic analysis of MLST data and generation of minimum spanning tree

The resulting set of FASTA sequences were concatenated and a phylogenetic tree was constructed by the unweighted-pair group method using average linkages (UPGMA) with 1,000 replicates for bootstrap testing (MEGA X and/or BioNumerics version 8.1.1). Each clade was identified by using the goeBURST version 1.2.1. Briefly, 4.175 DSTs were downloaded from the MLST database as of July 2022, then compared to those of CF isolates by using eBURST algorithm. The DSTs were divided according to an optimal solution for the placement of links between the different DSTs and the organization of strains into clonal clusters. Clades were assigned according to the cluster distribution and the placement of each founder strain of the different clonal complexes. To elaborate the population structure and link together the more closely related isolates originating from different CF patients, a Minimum spanning tree was generated based on the MLST data and the goeBurst cluster output. The analysis was performed using BioNumerics version 8.1.1 (Applied Maths).

Media and growth conditions

C. albicans isolates were routinely grown in YPD medium (1% yeast extract, 2% peptone, 2% dextrose) at 30°C under shaking, unless stated otherwise. Agar in a final concentration of 2% was supplemented for medium solidification. To test the ability of C. albicans clinical isolates to filament, colony morphologies were examined on YPD, YP (without dextrose), SD (6.7 g/L Yeast Nitrogen Base without amino acids and ammonium sulfate, 2% dextrose), Spider (1% peptone, 1% yeast extract, 1% manitol, 0.5% NaCl, and 0.2% K₂HPO₄) and RPMI 1640 + GlutaMAX (gibco) supplemented with 50 mM HEPES buffer (Sigma) agar media. Each clinical isolate was diluted to an optical density at 600 nm (OD_600nm) of 0.1, spotted onto the agar plate then incubated at 37°C for 3 to 5 days to assess colony morphology. Colonies were photographed with a 10× magnification using an EOS 550D camera (Canon, Tokyo, Japan) connected to a Leica M80 stereomicroscope (Leica Microsystems, Rueil-Malmaison, France).

Strain fitness assay and quantification of doubling times

To monitor for growth defects among CF isolates, growth curves were generated using a TECAN Sunrise device by measuring the OD_600nm every 10 min for 42 h (S3 Fig). Strains were individually grown three-times independently in 96-well plates at a starting OD_600nm of 0.1 in 100 μl of YPD or SD. The temperature was set at 30°C. TECAN OD_600nm readings were converted into “flask OD_600nm” reading using the following formula: OD_Flask = OD_Tecan × 12.2716–1.0543 [95] and doubling times were calculated within the exponential growth interval as previously described [96]. The resulting data were analyzed using GraphPad Prism 8.0.2.

C. albicans biofilm growth assays

Biofilm growth assays were performed using two methods, the standard optical density assay method [21] and the XTT assay method [97]. The standard optical density assay was performed in 24-well polystyrene plates. Plates were pretreated overnight with fetal bovine serum at 37°C (FBS, gibco, catalog # 10270–106). The following day, each well was inoculated with 2 ml of cells at an OD_600nm of 0.3 and incubated at 37°C for 90 min at 110 rpm to allow cell adherence. After 90 min, the plates were washed with 2 ml of 1X phosphate-buffered saline (PBS) to eliminate non-adhered cells and replaced with 2 ml of fresh YPD. Plates were sealed with breathseal sealing membranes (Greiner bio-one) and incubated at 37°C for 24 h with shaking at 110 rpm. The medium was aspirated carefully, and each well was washed twice with 1 ml of 1X PBS. The OD_600nm was read on a Tecan Spark multimode reader (Tecan Systems). Sixty-one measurement points at independent locations in each well of the 24-well plate were performed and a heatmap was generated for each well. The average density for each strain was calculated from three independent experiments after subtracting the background value of the control well with YPD only. Biofilm layer was also photographed using an EOS 550D camera (Canon, Tokyo, Japan).

The XTT assay method [97] was performed in FBS-precoated polystyrene 96-well microtiter plates (TPP 96-Well, Flat-Bottom Microplate) in YPD medium at 37°C then cell viability was assessed using the tetrazolium salt (2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino)carbonyl]-2H-tetrazolium hydroxide (XTT) (Sigma) colorimetric method to detect cellular metabolic activities within C. albicans biofilms. Strains were grown overnight in YPD at 30°C under shaking. The following day, 100 μl of cell suspension containing 10⁶ cells/ml in YPD medium were pipetted into wells of the 96-well flat-bottom microplate plates and incubated at 37°C for 90 min to allow cell adhesion on the surface of wells. After the incubation phase, YPD medium was aspirated off and wells were washed twice with PBS to eliminate non-adherent cells, then fresh YPD was added. The plates were covered with BREATHseal sealing membranes (Greiner bio-one) and incubated for 24 h. XTT solution was prepared at 0.5 g/L in PBS and filter sterilized with a 0.22 μm filter, aliquoted, and conserved at –70°C for later use. Prior to using, XTT aliquot was thawed out and a stock solution of 10 mM menadione (ChemCruz) was dispensed to a final concentration of 1 μM, then a volume of 100 μl aliquot of XTT/menadione was added to each pre-washed well. The plates were wrapped with aluminum foil as XTT is sensitive to light exposure and incubated for 1 h at 37°C. The colorimetric shift was read at 490 nm on a Tecan Spark multimode reader (Tecan Systems).

Bacterial strains used in this study

P. aeruginosa reference strain PAO1 (a gift from Laurent Debarbieux, Institut Pasteur, Paris, France) and clinical strain Pa29575 isolated from patient CF02 were used for dual-species biofilm assay. Prior to each experiment, the bacterial strains were revived from frozen stocks at -80°C by streaking on Luria Bertani (LB) plates and left for overnight incubation at 37°C. Bacterial colonies were then inoculated at 37°C and incubated on an orbital shaker at 200 rpm overnight with aeration prior to co-culture assays with C. albicans.

Dual-species biofilms of C. albicans and P. aeruginosa

The protocol for C. albicans-P. aeruginosa biofilm assay was kindly provided by Dr. Rebecca Hall. Assays were performed as described by Alam et al. [14,18,19] with slight modifications. Briefly, C. albicans was grown overnight in YPD and P. aeruginosa was grown overnight in LB, and cultures were washed in PBS. C. albicans strains were resuspended at 1 × 10⁶ cells/mL and P. aeruginosa strains were resuspended to an OD_600nm of 0.2 in Dulbecco Modified Eagle Medium (DMEM, Sigma) supplemented with 1% L-glutamine. Each well of 96-well plates contained 100 μL (1×10⁵ cells) C. albicans and 10 μL (2.5×10⁶ cells) of P. aeruginosa (final ratio fungus:bacterium = 1:25). Plates were incubated at 37°C for 2 h to allow cells to adhere, at which point the medium was replaced with fresh sterile medium, and plates incubated statically at 37°C for 24 h. Medium was replaced with 100 μL PBS containing 50 μg/mL DNase I and plates were incubated at 37°C for 1 h to degrade the extracellular matrix. Biofilms were detached from the plates by scraping, serially diluted, and plated onto YPD agar supplemented with 100 μg/mL tetracycline to determine viable cells of C. albicans (colony forming units, CFUs). Percent CFUs of C. albicans cells co-cultured with P. aeruginosa PAO1 or Pa29575 was calculated relative to biofilm growth of C. albicans cells alone. The experiments were carried out 5 times independently. Statistical analysis was performed using the Mann-Whitney test.

Whole genome sequencing, variant-calling and phylogenetic analyses

Genomic DNA was prepared according to the method by Amberg et al. [98] from a set of 18 C. albicans clinical isolates (HBJ series, patient CF02, S2 Fig). Whole genome sequencing libraries were prepared at the Biomics core facility [99] at Institut Pasteur, Paris, France, using the Illumina TruSeq™ DNA PCR-Free Low Throughput Library Prep Kit (FC-121-3001) following the manufacturer’s instructions. Libraries were sequenced on an Illumina NovaSeq 6000 device, generating 7 to 12 million 150-bp paired-end reads. Reads have been deposited at the NCBI Sequence Read Archive under BioProject ID PRJNA1008086.

Sequences and genomic variations were analysed essentially as described by Ropars et al. [20] and Marton et al. [100]. Briefly, each set of paired-end reads was mapped against the C. albicans reference genome SC5314 haplotype A or haplotype B [101] downloaded from the Candida Genome Database (version A22-s07-m01-r57) [102] using Minimap2 version 2.9 [103]. SAMtools, version 1.9, and Picard tools, version 2.8.1 (http://broadinstitute.github.io/picard), were then used to filter, sort, and convert SAM files. SNPs were called using the Genome Analysis Toolkit (GATK) version 3.6, according to GATK best practices [104]. SNPs were filtered using the following parameters: VariantFiltration, QD < 2.0, LowQD, ReadPosRankSum < −8.0, LowRankSum, FS > 60.0, HightFS, MQRankSum < −12.5, MQRankSum, MQ < 40.0, LowMQ, HaplotypeScore > 13.0, HaploScore. Coverages were also calculated using GATK. Besides passing GATK’s filters, we also checked for read depth (it had to be between 0.5 and 1.5 of the mean genome coverage). Highly confident SNPs were displayed on Integrative Genomics Viewer version 2.9.4. [105] by creating a bed file with color-coding of SNPs, enabling navigation on the C. albicans genome and examination of SNPs in candidate genes. The GATK variant filtration walker (VariantAnnotator) was used to add allele balance information to.vcf files. The value of allele balance at heterozygous sites (ABHet) is a number that varies between 0 and 1. ABHet is calculated as the number of reference reads from individuals with heterozygous genotypes divided by the total number of reads from such individuals. A diploid genome will be defined by an ABHet value of 0.5. A triploid strain may contain either three identical alleles (an allelic frequency of 1) or two identical alleles and one different allele (frequencies of 0.66 and 0.33). A tetraploid strain may have allelic frequencies of either 0.5 (2×2 identical alleles), 1 (4 identical alleles), or 0.25 and 0.75 (3 identical alleles and 1 different allele). Heterozygous SNP density maps were constructed as described by Loll-Krippleber et al. [106], by determining the number of heterozygous positions per 10-kb region and plotting each value using Python.

We used RAxML (Randomized Axelerated Maximum Likelihood), a tool for phylogenetic analysis and post-analysis of large phylogenies, to construct phylogenetic relationships between 200 C. albicans isolates based on highly confident SNPs identified in each of 182 strains from Ropars et al. [20] and the 18 isolates from patient CF02, with 1,000 bootstrap replicates. The final circular maximum likelihood phylogenetic tree output was generated using iTOL for the annotation and display of isolates that were color-coded by clade.

C. albicans transformation

C. albicans cells were transformed using the standard lithium acetate procedure, as described previously [107], with minor modifications. Briefly, an overnight culture of C. albicans was diluted to an OD_600nm of 0.2 in 50 mL of fresh YPD medium and incubated at 30°C under shaking until OD_600nm reached 0.7. The culture was centrifuged at 3,000 rpm for 5 min and the cell pellet was washed once in cold 10X Tris-EDTA (100 mM Tris pH 7.5, 10 mM EDTA). The cells were pelleted and resuspended in 750μL 1X LiAc buffer (100 mM lithium acetate, 10 mM Tris-HCl, 1 mM EDTA pH 7.5), then placed at room temperature for 30 minutes. One hundred microliters of competent cells were mixed with 1 to 10 μg of DNA, 50 μg of sonicated salmon sperm DNA (Sigma-Aldrich) and 700 μL of 1X LiAc supplemented with 40% polyethylene glycol 4000 (Sigma–Aldrich). The transformation mixture was then incubated overnight at 30°C. Heat-shock treatment was conducted at 44°C in a water bath for 15 minutes. Cells were gently resuspended in 3 mL YPD and incubated for 5 h to allow transformed cell recovery, prior to plating on selective media. Transformants that integrated the caSAT1 resistance markers were selected on YPD medium supplemented with 200 μg/mL nourseothricin (Jena Bioscience) and incubated at 30°C for two days. Transformants that integrated the hygB resistance marker were selected on YPD medium supplemented with 800 μg/mL hygromycin B (Sigma-Aldrich) and incubated at 30°C for two days.

C. albcians mutant strain construction

The C. albicans mutant strains used in this study are listed in S8 Table. Oligonucleotides used for mutant creation are listed in S3 Table. To generate homozygous ROB1 deletion mutants in clinical isolates HBJ6-2 and HBJ6-3, we used a transient CRISPR-Cas9 genome editing approach as described in Min et al. [30]. The CaCAS9 expression cassette (containing the ENO1 promoter, CaCAS9 open reading frame, and CYC1 terminator) was amplified from plasmid pV1093 [30] with primer pairs CaCas9/F and CaCas9/R (S3 Table), using DNA polymerase Ex Taq (TaKaRa, Japan). The sgRNA expression cassette containing the SNR52 promoter, guide sequence, and sgRNA scaffold sequence was assembled by the single-joint PCR method [108]. In the first step, the SNR52 promoter and sgRNA scaffold components were PCR amplified using flanking primers SNR52/F and sgRNA/R and internal chimeric primers SNR52/R_ROB1 and sgRNA/F_ROB1 (S3 Table). The chimeric primers overlapped by a 20-base segment that specified the guide sequence. In the second step, both components were joined by primer extension, relying upon annealing of the complementary chimeric primer extensions. In the third step, the joined product was PCR amplified with nested primers SNR52/N and sgRNA/N (S3 Table), to yield the sgRNA cassette. The repair template, carrying the caSAT1 dominant selection marker conferring resistance to nourseothricin, was amplified from plasmid pV1093 using oligos NAT_ROB1_repair/F and NAT_ROB1_repair/R (S3 Table). Yeast cells were co-transformed with 1 to 3 μg of each of the resulting PCR-amplified DNA fragments. Correct repair template cassette integration and the deletion of both alleles of ROB1 were confirmed by PCR, using primers flanking the expected repair template integration sites Verif-1-F, Verif-1-R, Verif-2-F, and Verif-1-R (S3 Table).

We used site-directed mutagenesis by fusion PCR to generate ROB1^G299E and ZFU3^A127T allele replacement cassettes and create ROB1/ROB1^G299E and ZFU3/ZFU3^A127T heterozygous mutants in clinical isolate HBJ6-2 (S8 Fig). The strategy relies on three fusion PCR reactions using four independently generated PCR products, namely ROB1-Frg1, ROB1-Frg2, caSAT1-Frg3 and Dw-Frg4 for the ROB1^G299E allele replacement cassette (S8 Fig, panels A and B); and ZFU3-Frg1, ZFU3-Frg2, hygB-Frg3 and Dw-Frg4 for the ZFU3^A127T allele replacement cassette (S8 Fig, panel C). All PCR reactions were performed using DNA polymerase Ex Taq (TaKaRa, Japan). For generating the ROB1/ROB1^G299E mutants (or respective control strains, where primers for site directed mutagenesis were not included in the design and were omitted from the PCR reactions), genomic DNA from strain HBJ6-2 and both primer pairs ROB1(I)F and ROB1(I)R (S3 Table, hybridize at positions bps 144 to 164 and bps 941 to 983 with respect to the ATG translation start site of ROB1, respectively) and ROB1(II)F and ROB1(II)R (S3 Table, hybridize at position bps 941 to 983 and bps 3,358 to 3,382 with resepct to the ATG translation start site of ROB1, respectively) were used to PCR-amplify fragments ROB1-Frg1 (840 bp, S8 Fig, panel A) and ROB1-Frg2 (2,465 bp, S8 Fig, panel A) that overlap by a 43-bp segment created by complementary primers ROB1(II)F and ROB1(I)R designed to respectively introduce G>A and C>T complementary mismatches at position 962 relative to the ATG translation start site of ROB1 (n.b. ROB1 carries an intron from position 48 to 113 with respect to the ATG translation start site). Fragment caSAT1-Frg3 (1,278 bp, S8 Fig, panel A), carrying the caSAT1 dominant nourseothricin resistance marker for the selection of integrative transformants, was generated using pFA6-SAT1 plasmid [109] as a template and chimeric primers SAT1F (S3 Table, overlapping with primer ROB1(II)R) and SAT1R (S3 Table, overlapping with primer ROB1 downstreamF, see below). Fragment Dw-Frg4 (412 bp, S8 Fig, panel B), carrying the downstream recombination region, was PCR-amplified from strain HBJ6-2 genomic DNA using primers ROB1 downstreamF and ROB1 downstreamR (S3 Table, hybridize at positions bp 3,383 to bp 3,400 and bp 3,755 to bp 3,774 with respect to the ATG translation start site of ROB1, respectively). Fragments ROB1-Frg1 and ROB1-Frg2 were fused by primer extension, relying upon annealing of the complementary 43-bp segment, to yield a 3,262-bp product carrying the G>A substitution in ROB1 and including a 340-bp sequence downstream of the TAA stop codon of ROB1 to serve as a putative transcription termination sequence (Fusion PCR #1). Fragments caSAT1-Frg3 and Dw-Frg4 were fused by primer extension, relying upon annealing of the complementary 38-bp segment brought by primers ROB1 downstreamF and SAT1R, to yield a 1,650-bp product carrying the caSAT1 selection marker and a 392-bp region allowing recombination at positions bps 341–732 downstream of the TAA stop codon of ROB1 (Fusion PCR #2). Fusion PCR products #1 and #2 were joined in a third fusion PCR relying upon annealing of the complementary 48-bp segment brought by chimeric primer sequences ROB1(II)R and SAT1F from fragments ROB1-Frg2 and caSAT1-Frg3, respectively, to yield the 4,866-bp ROB1^G299E allele replacement cassette (S8 Fig, panel D). The same strategy was used to generate the 4,777-bp ZFU3^A127T allele replacement cassette (and the equivalent wild-type ZFU3 allele replacement cassette as a control, S8 Fig, panel D). Briefly, DNA fragments ZFU3-Frg1 (738 bp, amplified with primers ZFU3(I)F and ZFU3(I)R, S3 Table), ZFU3-Frg2 (1,618 bp, amplified with primers ZFU3(II)F and ZFU3(II)R, S3 Table), hygB-Frg3 (2,002 bp, amplified with primers HygBF and HygBR, S3 Table) and Dw-Frg4 (568 bp, amplified with primers ZFU3 downstreamF and ZFU3 downstreamR, S3 Table) were PCR-amplified using genomic DNA from strain HBJ6-2 (for ZFU3-Frg1, ZFU3-Frg2 and Dw-Frg4) and plasmid pAU34-HygB [110] (for hygB-Frg3) as a template (S3 Table and S8 Fig, panel C). Fragments ZFU3-Frg1 and ZFU3-Frg2 were fused by primer extension, relying upon annealing of a complementary 46-bp segment brought by primers ZFU3(II)F and ZFU3(I)R, to yield a 2,310-bp product carrying the G>A substitution in ZFU3 at position bp 489 downstream of the ATG translation start site and including a 620-bp sequence downstream of the TGA stop codon of ZFU3 to serve as a putative transcription termination sequence (Fusion PCR #1). Fragments hygB-Frg3 and Dw-Frg4 were fused by primer extension, relying upon annealing of the complementary 49-bp segment brought by primers ZFU3-downstreamF and HygBR, to yield a 2,521-bp product carrying the hygB selection marker and a 543-bp region allowing recombination at positions bps 621–1,163 downstream of the TGA stop codon of ZFU3 (Fusion PCR #2). Fusion PCR products #1 and #2 were joined in a third fusion PCR relying upon annealing of the complementary 52-bp segment brought by chimeric primer sequences ZFU3(II)R and HygBF from ZFU3-Frg2 and hygB-Frg3 fragments, respectively, to yield the 4,777-bp ZFU3^A127T allele replacement cassette (S8 Fig, panel D). The correct integration of ROB1^G299E (S8 Fig, panel E) and ZFU3^A127 allele (S8 Fig, panel F) replacement cassettes was confirmed with primer pairs ROB1-Veri-cassetteF and ROB1-Veri-cassetteR and ZFU3-Veri-cassetteF and ZFU3-Veri-cassetteR (S3 Table). The presence of the G>A heterozygous substitutions at positions bp 962 and bp 489 downstream of the ATG translation start sites of ROB1 and ZFU3 in the resulting HBJ6-2 mutants, respectively, was confirmed via Sanger sequencing (S8 Fig, panel G) using primer pairs ROB1-Veri-MutationF and ROB1-Veri-MutationR and ZFU3-Veri-MutationF and ZFU3-Veri-MutationR, respectively (S3 Table).

Gene expression analyses by RT-PCR and qRT-PCR

The HBJ6-2 strain derivatives ROB1^G299E-SAT1/ROB1 and ROB1-SAT1/ROB1 were grown overnight in liquid YPD at 30°C. The following day, cells were resuspended at an OD_600nm of 0.2 in liquid YPD medium and regrown for four hours under vigorous shaking at 30°C. The cultures were centrifuged at 3,000 rpm for 5 min, and the cell pellets were used for total RNA isolation with the RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions for yeast and filamentous fungi. cDNA synthesis from the extracted RNA was performed using the QuantiTect Reverse Transcription Kit (Qiagen). For each reaction, equal amounts of RNA were used (1 μg in a 20-μl reaction mixture). The expression of each gene (ACT1, TEF3, ROB1, ALS3, HWP1 and ECE1) was inspected by PCR amplification (thirty cycles with 15 seconds at 95°C, 15 seconds at 50°C and 40 seconds at 70°C using Taq DNA polymerase) from the synthesized cDNA with gene-specific primers listed in S3 Table, followed by agarose gel electrophoresis of the PCR products (Fig 6A). Quantitative real-time PCR (Fig 6B) was performed in an Applied Biosystems QuantStudio 3 Real-Time PCR System, in a 25-μl reaction mixture containing 1X SYBR Green PCR Master Mix (Applied Biosystems, catalog # 4309155), 100 nM of both forward and reverse primers (see S3 Table for qPCR primers used in this study), and 1 μl of cDNA. To quantify the relative expression of the ROB1, ALS3, HWP1, and ECE1 genes, the expression of endogenous gene ACT1 was used as the normalization standard, and the relative expression of each gene in the ROB1^G299E-SAT1/ROB1 mutant compared to its expression in the ROB1-SAT1/ROB1 was determined using the cycle threshold (ΔΔCt) method using the average ΔCt values of the ACT1 gene in the ROB1-SAT1/ROB1 strain as a calibrator. The assays were carried out three times independently, and statistical analysis was performed using a Welch’s t-test.